Structural and Mechanistic Studies on ThiO, a Glycine Oxidase

IscS is a PLP-utilizing enzyme that reacts with cysteine to give an active site persulfide (2). This persulfide adds to the carboxy-terminal acyladeny...
0 downloads 0 Views 629KB Size
Download PDF
Biochemistry 2003, 42, 2971-2981

2971

Structural and Mechanistic Studies on ThiO, a Glycine Oxidase Essential for Thiamin Biosynthesis in Bacillus subtilis†,‡ Ethan C. Settembre, Pieter C. Dorrestein, Joo-Heon Park, Amy M. Augustine, Tadhg P. Begley,* and Steven E. Ealick* Department of Chemistry and Chemical Biology, Cornell UniVersity, Ithaca, New York 14853 ReceiVed September 27, 2002; ReVised Manuscript ReceiVed January 22, 2003

ABSTRACT:

The thiO gene of Bacillus subtilis encodes an FAD-dependent glycine oxidase. This enzyme is a homotetramer with a monomer molecular mass of 42 kDa. In this paper, we demonstrate that ThiO is required for the biosynthesis of the thiazole moiety of thiamin pyrophosphate and describe the structure of the enzyme with N-acetylglycine bound at the active site. The closest structural relatives of ThiO are sarcosine oxidase and D-amino acid oxidase. The ThiO structure, as well as the observation that N-cyclopropylglycine is a good substrate, supports a hydride transfer mechanism for the enzyme. A mechanistic proposal for the role of ThiO in thiazole biosynthesis is also described.

Thiamin pyrophosphate 13 plays a key role in the stabilization of acyl carbanion intermediates and is an essential cofactor in all living systems. The biosynthesis of thiamin is complex and is not yet well-understood. The thiazole and the pyrimidine moieties are synthesized in separate branches of the pathway and are coupled to form thiamin phosphate, which is phosphorylated to generate the active form of the cofactor. In some bacteria (e.g., Escherichia coli), the thiazole moiety is synthesized from cysteine, deoxy-D-xylulose 5-phosphate, and tyrosine in a set of reactions requiring ThiF, ThiS, ThiG, ThiH, ThiI, and IscS. In other bacteria (e.g., Bacillus subtilis), the thiazole moiety is synthesized from cysteine, deoxy-D-xylulose 5-phosphate, and glycine, and sequence analysis suggests that the biosynthesis is catalyzed by ThiF, ThiS, ThiG, ThiO, ThiI, and NifS (Figure 1) (1-3). The specific reactions catalyzed by several of the thiazole biosynthetic proteins have now been identified. IscS is a PLPutilizing enzyme that reacts with cysteine to give an active site persulfide (2). This persulfide adds to the carboxyterminal acyladenylate of ThiS, and the resulting acyl disulfide undergoes a disulfide exchange reaction with Cys184 of ThiF to form a covalent ThiF-ThiS complex linked by an acyl disulfide linkage. This intermediate is likely to be the sulfur donor for the thiazole biosynthesis in E. coli (4). In addition to its role in the formation of the acyl disulfide, ThiF also catalyzes the formation of the acyl adenylate of ThiS (5). The specific reactions catalyzed by ThiG and ThiH have not yet been identified. † This work was supported by National Institutes of Health Grant DK44083 (to T.P.B.). S.E.E. is indebted to the W. M. Keck Foundation and the Lucille P. Markey Charitable Trust. ‡ The Protein Data Bank entry for ThiO is 1NG4 and for the ThiO complex with N-acetylglycine is 1NG3. * To whom correspondence should be addressed: Department of Chemistry and Chemical Biology, Cornell University, Ithaca, NY 14850. Telephone: (607) 255-7961. Fax: (607) 255-1227. E-mail: [email protected] or [email protected].

In B. subtilis, ThiS thiocarboxylate 3 rather than the ThiFThiS acyl disulfide is likely to be the immediate sulfur donor (TPB unpublished) and ThiF, ThiS, and ThiG are likely to have functions identical or very similar to those of the E. coli enzymes. However, ThiO and ThiH share no sequence similarity. These differences in substrate utilization and enzymatic requirements suggest subtle differences between the biosynthetic strategies for thiazole formation in E. coli and B. subtilis. However, these differences remain to be elucidated because thiazole biosynthesis has not yet been reconstituted in a cell free system. ThiO (YjbR) was previously isolated as a homotetramer of 42 kDa subunits and shown to be an FAD-dependent glycine oxidase (6). Amine oxidases have been extensively studied, and the structures of sarcosine oxidase (7), D-amino acid oxidase (8, 9), L-aspartate oxidase (10), and polyamine oxidase (11) have been reported. The mechanism for the reaction is still controversial, and mechanisms involving an electron transfer, a hydride transfer, and a covalent amine flavin adduct are all still considered possible. Sequence homology searches indicated that ThiO is 18-27% identical with the sarcosine oxidases and D-amino acid oxidases (3, 12). These proteins are characterized by a conserved amino acid sequence of the FAD binding glutathione reductase 2 (GR2)1 family (13). In this paper, we demonstrate that the glycine oxidase encoded by the thiO gene is essential for thiazole formation in B. subtilis, we describe mechanistic and structural studies on this enzyme, and we propose a role for ThiO in the biosynthesis of the thiazole moiety of thiamin. 1 Abbreviations: SOX, sarcosine oxidase; DAAOX, D-amino acid oxidase; pkDAAOX, pig kidney D-amino acid oxidase; rgDAAOX, Rhodotorula gracilis D-amino acid oxidase; CIP, calf intestinal alkaline phosphatase; GR2, glutathione reductase 2; PEG 2K MME, polyethylene glycol monomethyl ether 2000; CHESS, Cornell High Energy Synchrotron Source; LB, Luria-Bertani medium; DTT, dithiothreitol; APS, Advanced Photon Source; DXP, deoxy-D-xylulose 5-phosphate; FAD, flavin adenine dinucleotide; SeMet, selenomethionine.

10.1021/bi026916v CCC: $25.00 © 2003 American Chemical Society Published on Web 02/22/2003

2972 Biochemistry, Vol. 42, No. 10, 2003

Settembre et al.

FIGURE 1: Mechanistic proposal for the role of ThiO in the biosynthesis of the thiazole moiety of thiamin pyrophosphate in B. subtilis.

MATERIALS AND METHODS Chemical Reagents. Thiazole, glycine, cysteine, horseradish peroxidase, L-alanine [99% enantiomeric excess (ee)], D-alanine (99% ee), 1-amino-2-propanol, (S)-(+)-2-amino1-propanol (97% ee), (R)-(+)-2-amiopropanol (97% ee), cyclopropylamine, and 2,2-dideuterioglycine (98% deuterated) were from Sigma-Aldrich. [2-13C]Glycine was from Cambridge Isotope Laboratories, and phenol was from Mallinckrodt. Growth CurVes for ThiO- B. subtilis. The ThiO deficient B. subtilis was a gift from J. Errington (Oxford University, Oxford, England) (14). The fastest growing colonies on LB/ erythromycin plates were used to inoculate 3 mL of minimal medium. After 24 h at 37 °C, the slowest growing culture was used to inoculate 50 mL of minimal medium. This culture was maintained at 37 °C for 24 h to starve the cells for thiamin and divided into two 25 mL cultures, and an additional 25 mL of minimal medium was added to each. Thiazole alcohol (2 µM) was added to one of the cultures. The absorbance at 595 nm was monitored to obtain the growth curves. Special precautions were taken to ensure that the flasks used for the growth curves had not been exposed to full medium, thiamin, or thiazole. One liter of minimal medium used for the growth curve contained 200 mL of 5× minimal salt solution, 20 mL of 20% glucose, 5 mL of 1% L-tryptophan, 50 mL of 4% L-glutamine, 2 mL of 2 mg/mL FeCl3, 2 mL of 0.1 mg/mL MnSO4, and 10 mL of the 100× trace elements. The 5× minimal salt solution contained 0.057 M K2SO4, 0.31 M K2HPO4‚5H2O, 0.017 M sodium citrate, and 0.004 M MgSO4 and was adjusted to pH 7 with 10 N NaOH. The 100× trace elements contained 0.55 g of CaCl2, 0.17 g of ZnCl2, 0.043 g of CuCl2‚2H2O, 0.06 g of CoCl2‚ 6H2O, and 0.06 g of Na2MoO4‚2H2O in 1 L of solution. ThiO OVerexpression. The thiO gene was PCR amplified from B. subtilis genomic DNA (strain CU1065) using the following primer pair: 5′-GCT AAA GGA GAT GCC ATA TGA AAA GGC ATT ATG AAG C-3′ (inserts an NdeI site at the start codon) and 5′-CGT CTT TAC CGT TCA GCT CGA GCA TCA TAT CTG AAC CGC C-3′ (inserts an XhoI site after the stop codon). The PCR product was completely

digested with NdeI and XhoI, resulting in two fragments. One fragment, from an internal NdeI site to the engineered XhoI site, was ligated into similarly digested pET-16b to give pCLK810. This plasmid was digested with NdeI and dephosphorylated with CIP. The other fragment from the PCR product digest, from the engineered NdeI site to the internal NdeI site, was ligated into the cut and dephosphorylated pCLK810. Clones were screened for the direction of the insertion, and one with the correct orientation was named pCLK811 (thiO in pET-16b). Protein Purification. The native protein was obtained by inoculating 1 L of LB and 100 µg/mL ampicillin with 5 mL of a saturated starter culture. The cells were grown at 37 °C until they reached an OD600 of ∼0.6, at which point the temperature was lowered to 30 °C and the cells were induced with 500 µM isopropyl β-D-thiogalactoside. After induction for 4 h, the cells were spun down at 5000 rpm for 10 min and stored at -80 °C. For production of the protein with selenomethione (SeMet) incorporated, the methionine auxotrophic strain of E. coli, B834(DE3) (Novagen), was transformed with pCLK811 and the cells were grown using growing conditions slightly varied from those described above. The 1 L of growth medium contained M9 salts supplemented with all amino acids (40 µg/mL each) except L-methionine, which was replaced with L-SeMet, 0.4% (w/v) glucose, 2 mM MgSO4, 25 µg/mL FeSO4‚7H2O, 1 mM CaCl2, 100 µg/mL ampicillin, 1 mM DTT, and a 1% BME vitamin solution (GibcoBRL). Also, the cells from the initial 5 mL starter culture were washed with the above medium, and used to start a 50 mL culture. This second culture was grown to an OD600 of ∼0.6 and used to inoculate the larger 1 L culture. The rest of the expression was performed as described above. All purification steps were carried out at 4 °C. All buffers contained 1 mM DTT for the protein with SeMet incorporated. The cells were resuspended in 20 mL of wash buffer [50 mM NaH2PO4, 300 mM NaCl, and 20 mM imidazole (pH 8.0)] and broken using a French press. The crude extract was centrifuged, and the resulting supernatant was mixed for 1 h with 500 µL of Ni-NTA beads (Novagen) equilibrated

X-ray Structure and Mechanism of ThiO with the wash buffer. The beads were then added to a polypropylene column and were washed with 300 mL of wash buffer. The column was first eluted with 20 mL of wash buffer containing 50 mM imidazole to remove weakly binding proteins. ThiO was eluted from the column using wash buffer containing 150 mM imidazole. After gel filtration using an Econo-Pac 10DG column (Bio-Rad) with elution with 50 mM Tris (pH 7.0), the protein was concentrated to 10 mg/mL using a 10 kDa cutoff concentrator (Amicon) and stored at -80 °C. Protein concentrations were determined by the Bradford method (15) using bovine serum albumin as the standard. The purity of ThiO was determined by Coomassie-stained SDS-PAGE and found to be greater than 99% (data not shown). Synthesis of Cyclopropylglycine (15). Chloroacetic acid (50 mg, 0.53 mmol) was added at room temperature with stirring to cyclopropylamine (3 g, 53 mmol). After 3 days, the unreacted cyclopropylamine was removed by vacuum distillation. The resulting crude N-cyclopropylglycine was dissolved in methanol and then rotary evaporated to remove residual cyclopropylamine. The product was purified by recrystallization from a 33.9:33:33:0.1 chloroform/ethyl acetate/acetone/ethanol mixture, to give 15 as a white crystalline solid (37 mg, 61%): 1H NMR (400 MHz, D2O) δ 3.59 (2 H, s), 2.64 (1 H, m), 0.66 (4 H, m); 1H NMR (400 MHz, 10 mM phosphate buffer in D2O, pD 7.4) δ 3.49 (2 H, s), 2.54 (1 H, m), 0.15 (4 H, m); EI-MS (M+) 116.1; EI-MS-MS (M+) 116.1, 88.1, 70.2. Synthesis of Compounds 16-18. These were prepared in a manner similar to that of compound 15. Compound 16: 1 H NMR (D2O, rotamer mixture) δ 3.55 (dd, 1 H), 3.35 (m, 3 H), 3.05 and 3.20 (sextet, 1 H), 1.01 and 1.04 (d, 3 H); EI-MS (M+) 134.1; EI-MS-MS (M+) 134.1, 116.1, 88.1, 70.2. Compound 17: 1H NMR (D2O, rotamer mixture) δ 3.55 (dd, 1 H), 3.35 (m, 3 H), 3.05 and 3.20 (sextet, 1 H), 1.01 and 1.04 (d, 3 H); EI-MS (M+) 134.0; EI-MS-MS (M+) 134.0, 116.1, 88.1. Compound 18: EI-MS (M+) 134.0; EIMS-MS (M+) 134.0, 116.1, 88.1.

Enzyme Assays. ThiO activity was measured by monitoring hydrogen peroxide production as previously described (6). A typical reaction mixture contained 50 mM 4-aminoan-

Biochemistry, Vol. 42, No. 10, 2003 2973 tipyrine, 5 units of horseradish peroxidase, 2 mM phenol, and 8 mM glycine or glycine analogue (compounds 15-18) in 0.5 mL of 25 mM Tris-HCl (pH 7.8). The reaction was initiated by the addition of 25 µL of ThiO (8-12 mg/mL) and monitored over a period of 10 min using the increasing absorbance at 500 nm ( ) 5560 M-1 cm-1). For kinetic characterization, various concentrations of glycine or its analogues were incubated with the enzyme and the initial rates were fitted to the Michaelis-Menten equation using Sigmaplot. Characterization of the Glycine Oxidation Product. A solution of freshly prepared ThiO (0.35 mL, 3.4 mg/mL) and 2000 units of catalase were added to 4 mL of glycine or [2-13C]glycine (10 mM) in 40 mM KPi (pH 7.9). After 3 h, the reaction mixture was lyophilized, redissolved in D2O (1 mL), centrifuged, and analyzed by NMR (400 MHz). Characterization of the N-Cyclopropylglycine Oxidation Products. A stock solution (10 mM) of N-cyclopropylglycine in deuterated 10 mM phosphate buffer (pD 8.2) was prepared. ThiO [75 µL, 18 mg/mL in 20 mM phosphate buffer (pH 7.8)] was added to 0.75 mL of this stock solution, and the reaction was monitored using NMR with solvent suppression after 0, 3, 18, 36, and 72 h, at which time 75% of 15 was converted to formate, glyoxalate, and cyclopropylamine. Ring-opened products were not observed. Compound 15 was stable under these conditions in the absence of the enzyme. Kinetics of FlaVin Reduction. Glycine or deuterated glycine [150 µL, 400 mM in 200 mM Tris-HCl (pH 7.6) degassed by sonication under vacuum] was mixed 1:1 with ThiO (degassed by soniation under vacuum) (3.6 mg/mL) using a stopped flow instrument, and the absorbance decrease at 455 nm was recorded for 2-10 s at 0.002 s intervals. The stopped flow instrument was a Hi-Tech Scientific preparative Quench and stopped flow SHU PQ/SF-53 model with a Hi-Tech Scientific SU-40 spectrometer unit and Keithley Driverlinx software. The observation cell has a high-efficiency mixer with an optical path length of 2 mm and two emission windows (2 mm × 2 mm). The average dead time for this instrument is 16-20 ms. The kinetics were determined at room temperature. The data for each experiment were imported into Sigmaplot and fitted to a first-order exponential decay. Stereochemistry of the ThiO-Catalyzed Oxidation. LAlanine and D-alanine were tested as substrates using the hydrogen peroxide assay described above. Each reaction mixture contained 2 units/mL horseradish peroxidase, 100 mM 4-aminoantipyrine, 4 mM phenol, and 3-800 mM D-alanine or L-alanine in 500 µL. D-Alanine was a substrate, while L-alanine was not. To obtain an approximate relative rate for the two isomers, reaction mixtures containing 100 mM samples of L-alanine and D-alanine were incubated with ThiO and the hydrogen peroxide assay system for 4 h. No L-alanine oxidation was observed, giving kD/kL relative rates of g5000. Because the L-alanine sample was contaminated with 0.5% D-alanine, the rates were recorded after an initial incubation of 300 s to ensure removal of this contaminant. Crystallization of ThiO. ThiO was crystallized using the hanging drop method with each drop containing 2 µL of protein and 1 µL of well solution. The protein concentration was 10 mg/mL, and the well solution for optimized conditions contained 18% polyethylene glycol monomethyl ether 2000 (PEG 2K MME) and 100 mM 4-(2-hydroxyethyl)-1-

2974 Biochemistry, Vol. 42, No. 10, 2003 piperazineethanesulfonic acid (pH 7.15). Crystals appeared within 1 week and grew to their maximum size (0.5 mm × 0.5 mm × 0.35 mm) in 2 weeks. Crystals were yellow, indicative of the presence of the oxidized FAD cofactor. Preliminary X-ray analysis showed that the crystals belong to space group P6122 or P6522 with the following unit cell dimensions: a ) 139.39 Å and c ) 209.64 Å. The crystals contain two monomers per asymmetric unit, corresponding to a solvent content of 60%. The crystallization conditions for the native protein and protein with SeMet incorporated were essentially the same except that 1 mM DTT was added for the SeMet protein. In general, crystals of the SeMet protein were ∼0.1 mm smaller in each direction. Crystals of the N-acetylglycine complex were prepared by adding 5 mM N-acetylglycine to the crystallization solution and 10 mM N-acetylglycine during cryoprotection. N-Acetylglycine is not a substrate and inhibits the enzyme. For cryoprotection, the crystals were gently transferred to a stabilization solution that was similar to the mother liquor but with 30% PEG 2K MME. The crystals were then transferred into solutions with increasing ethylene glycol concentrations (1% steps until the final concentration of 11% was reached), frozen by plunging them into liquid nitrogen, and stored for later use. For the N-acetylglycine complex, 10 mM N-acetylglycine was present as well. X-ray Data Collection and Processing. A two-wavelength data set was collected at the F2 station of the Cornell High Energy Synchrotron Source (CHESS) to 3.25 Å resolution. Following calibration of the beam using a Se foil, a fluorescence scan was taken on a SeMet-ThiO crystal. For data collection, one wavelength was chosen at the maximum of f ′ (edge) and the other was selected at the maximum of f ′′ (peak). The remote wavelength data were not collected because of crystal decay. Data were collected over 75° using 90 s for each 1° oscillation with a crystal to Quantum 4 CCD detector (Area Detector Systems Corp.) distance of 250 mm. Bijvoet pairs were measured after each 10° wedge using inverse beam geometry. Throughout data collection, wavelength calibration was checked and the camera was adjusted to maximize the flux as needed. The DENZO suite of programs was used for integration and scaling of the data (16). The inverse and direct beams were initially scaled together before the separate wavelengths were scaled together. Single-wavelength data were taken at beamline 8-BM at the Advanced Photon Source (APS) using a Quantum 315 detector (Area Detector Systems Corp.) in binned format. Data were collected over a range of 120° using 20 s for each 0.5° oscillation at a crystal to detector distance of 340 mm. MOSFLM (17) was used for data integration, and SCALA (18) was used for scaling. All data collection statistics are summarized in Table 1. Structure Determination. The initial Se atom positions were determined utilizing shake-and-bake direct methods (19) as implemented in SnB (20). The peak wavelength data were used to calculate normalized anomalous differences (∆E) using the DREAR (21) suite of programs. A total of 2000 ∆E’s with ∆E/σ (∆E) values of >3.5 and 20 000 triplet invariants were used to carry out 1000 random trials with 80 cycles of phase refinement per trial. Five of these trials produced solutions as judged by the behavior of the shake-

Settembre et al. Table 1: Summary of Data Collection and Processing Statistics Se (CHESS F2) wavelength (Å) resolution (Å) no. of reflections no. of unique reflections redundancy completenessa Rsym (%)a,b I/σa

(APS 8BM)

edge

peak

native

complex

0.9793 3.25 281836 19316 14.6 97.9 (96.7) 10.3 (38.6) 7.6 (3.2)

0.9791 3.25 291076 19189 15.2 97.6 (97.1) 10.3 (32.2) 8.0 (3.4)

0.9791 2.3 647137 51731 12.4 96.7 (96.7) 9.4 (34.9) 20.7 (6.2)

0.9791 2.6 164553 34445 4.6 96.4 (96.4) 6.8 (31.1) 17.9 (4.6)

a Values for the outer resolution shell are given in parentheses. Cornell High Energy Synchrotron Source (CHESS), MAD F2 data, 3.25 Å for edge and peak. Advanced Photon Source (APS) 8BM native ThiO data to 2.3 Å and N-acetylglycine complex data to 2.6 Å. b Rsym ) ∑∑i|Ii - 〈I〉|/∑〈I〉, where 〈I〉 is the mean intensity of N reflections with intensities Ii and common indices h, k, and l.

and-bake minimal function. From these phases, 18 of 24 expected Se atom positions were identified. These sites showed a noncrystallographic 2-fold axis perpendicular to a crystallographic 2-fold axis. The Se atom positions were input into CNS (22) and used for phasing. Data from both wavelengths were scaled together using CNS with the peak wavelength used as the reference. The initial phases were used to identify the six remaining Se atoms from a combination of anomalous difference Fourier and log-likelihood Fourier maps, and phases were recalculated using the 24 Se atom positions. Model Building and Structure Refinement. All model building was performed using the computer program O (23). Electron density maps were calculated for each possible space group, P6122 or P6522, and only the electron density for space group P6122 showed features consistent with a protein structure. The map was further improved with the addition of the noncrystallographic symmetry (NCS) constraints and a protein mask that was created from a bones representation of the electron density using the program MAPMAN (24). The backbone was traced for residues 5-53, 60-181, and 197-360, and the second monomer was generated using NCS. At this stage, FAD molecules were clearly visible in the electron density of each monomer and were included in the model. Combining calculated phases with the initial phases improved the map in all regions so that the entire chain could be traced from residue 1 to 360 and side chains were included for each monomer. The refinement procedure involved successive rounds of rigid body refinement, simulated annealing refinement, temperature factor refinement, and model rebuilding. In the native structure, some electron density for which we cannot account appeared above the isoalloxazine ring roughly in the position where the carboxyl group of Nacetylglycine was positioned in the complex. This density appeared to be too spread out to be a water molecule but not sufficiently spread out to be a series of water molecules. The density was modeled as a hydrogen peroxide molecule, one of the expected reaction products. The final refinement statistics are shown in Table 2. RESULTS ThiO- Mutant. The ThiO- mutant grown on minimal medium has an absolute requirement for the thiazole moiety of thiamin as illustrated in Figure 2.

X-ray Structure and Mechanism of ThiO

Biochemistry, Vol. 42, No. 10, 2003 2975

Table 2: Refinement Statistics and Model Building resolution (Å) total no. of non-hydrogen atoms no. of protein atoms no. of water oxygen atoms no. of ligand atoms no. of reflections in refinement no. of reflections in the test set R factora (%) Rfreeb (%) rms deviation from ideal geometry bonds (Å) angles (deg) Ramachandran plot most favored region (%) additional allowed region (%) generously allowed region (%) disallowed region (%) average B factor (Å2) main chain side chain water phosphate ligand FAD

Table 3: Glycine Analogues as Alternative Substrates for ThiO

ThiO

ThiO-N-acetylglycine

compound

kapp (min-1)

Km(app) (mM)

50-2.3 6008 5688 205 115 53499 4375 21.5 24.8

50-2.6 5924 5688 109 127 35656 1905 20.9 23.9

glycine 15 16 17 18 D-alanine L-alanine

1.2 ( 0.09 1.0 ( 0.07 poor activity at 25 mM poor activity at 25 mM 1.14 ( 0.06 0.40 ( 0.01 no activity (